One scan in · A repaired print out · No Photoshop needed
Upload one scan of a damaged print and repair it free online — creases, scratches, tears and fading gone, every face left exactly as it was, and the credit cost shown before you generate.
Creases, white scratch lines, torn corners, missing emulsion and brown foxing stains are rebuilt — the kinds of damage a sharpening slider cannot touch.
Faces, expressions, poses, clothing and background are left untouched. The print gets repaired; the moment is never reimagined.
It runs in the browser with free credits to start — no Photoshop, no layer masks, no clone-stamp work.
It is a tool that takes one scan of a damaged print and repairs the physical damage on it — the crease across the middle, the white scratch lines, the chipped corners, the brown foxing stains, the yellow-orange cast that decades of light left behind — while leaving the picture itself completely alone. This is different from sharpening.
A blurry digital snapshot needs detail invented; a damaged print needs damage removed and colour put back, and those are opposite jobs. Traditionally the work meant hours in Photoshop with the clone-stamp and healing brush, layer by layer, or paying a retoucher forty dollars and waiting a week per image. Here you upload the scan, the model reads which marks are damage and which are the photograph, and it rebuilds only the damaged parts in one pass.
On Kavel it runs on the Nano Banana image model through the KIE runtime, so a repaired print comes back in seconds rather than days. Start from your scan — the flow defaults to image-to-image — set the format and resolution, and the credit estimate updates before you commit; a failed run is never charged. The rule the tool follows matters more than the speed: your grandmother's face must come back as her face. Restoration that quietly redraws a jawline or smooths a smile has not restored anything, it has replaced it, and for the pictures people care about that is worse than leaving the crease alone.
If your scan is intact but soft or low-resolution, the HD photo converter is the better fit — that tool recovers sharpness, this one repairs damage. Photograph or scan the print flat and evenly lit and the model has the most of the original to work from.
Removes creases, scratches, tears, dust, foxing stains and missing emulsion — physical damage, not softness.
Corrects the yellow-brown cast of aged prints back to natural skin tones and recovers contrast in faded areas.
Leaves faces, poses, clothing and composition untouched, so the repaired print is the same moment, undamaged. Old photo damage and softness are different problems: the HD photo converter sharpens a picture that is intact but blurry, and the model comparison shows how the two engines behind old photo work differ on one prompt.
Creative engine
Upload a scan, confirm what needs repairing, choose the format and resolution, and the credit estimate updates before you generate.
A real edit made on Kavel with the same model this tool uses — a 1960s family snapshot with a deep crease, scratches, torn corners and heavy yellowing. Drag the divider and watch what does not move: four faces, four poses, the house, the path. Only the damage goes. Every output is AI-generated, so your result varies with your scan and the damage on it — treat this as a guide to what the tool can do, not a fixed result you will get back exactly.


Drag to compare before and after
From a creased scan to a clean print, with the cost shown before you commit.
Lay the print flat under even light and capture the whole frame. A straight, evenly lit scan gives the model the most of the original to work from — avoid angled phone shots with glare across the surface.
Name the damage you can see — a crease down the centre, water stains in one corner, a missing edge, colour that has gone orange. Specific direction repairs more than a generic request.
Confirm the estimate, generate the repair, and download at full resolution for printing. If a run fails it is free, so a second attempt with clearer direction costs nothing.
Different pictures, the same repair, one simple workflow.
Bring a box of creased, faded prints of parents and grandparents back to something you can share with relatives or print again for the wall.
Repair the one surviving picture of someone who has passed, so the version the family keeps is not the damaged one.
Clean up scans for a family tree, a heritage project, or an archive where the damage is getting in the way of the record.
Common questions about repairing damaged and faded prints with Kavel.
No, and the difference matters. Sharpening invents detail that a soft digital file never captured. This repairs damage on a print that was fine when it was taken — creases, scratches, tears, stains, fading. If your scan is undamaged but soft or low-resolution, use the HD photo converter instead; it is built for that job.
It should not, and that is the rule the tool works to. Faces, expressions, poses, clothing and background are left as they are; only the damage is rebuilt. If a repair does come back with a face that looks subtly different, retry and say explicitly to keep the faces unchanged — a failed or unwanted run costs nothing to redo.
No. This runs entirely in the browser with free credits to start, so there is no clone-stamp work, no layer masks and nothing to install. That is the main reason people arrive here after searching for a Photoshop tutorial — the manual method works, it just takes hours per image.
Creases, scratches, dust, foxing, torn corners, water stains and heavy colour fading all repair well. Where a large area of the picture is physically missing — a face torn away entirely — the model has to invent what was there, and inventing a face is exactly what you do not want. Crop to what survives instead.
Repair and colourisation are separate jobs. This tool restores what the print originally held, including recovering colour that has faded away. Adding colour that was never captured is a different request and a different result; ask for it explicitly if that is what you want.
Flat, square to the camera or scanner, evenly lit, and framed to include the whole print including its border. Glare across the surface and a steep angle both cost the model information it cannot recover. A flatbed scan beats a phone photo, but a careful phone photo works.
The credit estimate appears before you generate and updates with the format and resolution you pick, so you always see the cost first. Failed runs are never charged. New accounts start with free credits, which is enough to repair a print and judge the result before spending anything.
Yes — reprinting for a frame or a family album is one of the most common uses. Generate at the highest resolution available so the file holds up at print size, and keep your original scan; the repair is a new file, not a replacement for the archive copy.
One credit pool covers Nano Banana images and Seedance and Veo video. The cost shows before every run, and failed jobs are not charged. Use a subscription for ongoing work, or a one-time pack when you just need to top up.
What's included
What's included
What's included
One-time top-ups — buy extra credits any time you run low.
What's included
What's included
What's included
Powered by
Use one balance across the supported image and video models (Nano Banana 2 / Nano Banana Pro / Seedance 2 / Veo 3.1), and check the credit cost before each request.
Video models
Image models
Model credit guide
You're only charged for successful generations. The exact estimate in the generator varies by model, length, resolution, audio, and number of images.
| Type | Model | Credit cost |
|---|---|---|
| Video | Seedance 2.0 | 5s 720p image-to-video ≈ 188 credits; text-to-video ≈ 308 credits. Scales with resolution and length. |
| Video | Seedance 2 Fast | Faster and lower cost. 5s 720p text-to-video ≈ 248 credits. Mini is cheaper still. |
| Video | Veo 3.1 | Billed per video (Lite). About 34 credits at 1080p, about 23 at 720p. |
| Image | Nano Banana 2 | Generate or edit from text and images. About 8 credits per 1K image, 12 at 2K, 18 at 4K. |
| Image | Nano Banana Pro | Consistent run times across generations. About 12 credits per 1K or 2K image, 21 at 4K. |
| Image | Nano Banana 2 Lite | Faster, simpler variant. About 8 credits per image. Start here to test. |
Upload a scan, say what needs fixing, check the credits, and get the damage removed in one pass.